|
Back to 2014 Annual Meeting Posters
The Validity of Nonalcoholic Fatty Liver Disease Diagnosis in Veterans Affairs Administrative Databases
Srinevas K. Reddy*1,2, Willliam J. Culpepper2
1Surgery, University of Maryland School of Medicine, Baltimore, MD; 2Neurology, Baltimore Veterans Affairs Medical Center, Baltimore, MD
Introduction: Administrative data is often used to analyze clinical outcomes among large patient cohorts with common disorders. Yet few studies have defined the validity of the nonalcoholic fatty liver disease (NAFLD) diagnosis in administrative data despite the fact that NAFLD is the most common chronic liver disease in the United States. The aim of this study is to compare the accuracy of the NAFLD diagnosis without the presence of other chronic liver diseases using International Classification of Diseases-9 (ICD-9) codes in administrative data with that obtained from medical record review using Veterans Affairs (VA) sources.
Methods: VA Patient Treatment File and Outpatient Care File databases from 2010-2011 were searched for patients assigned a discharge or encounter ICD-9 diagnosis code of 571.8 (other chronic nonalcoholic liver disease). Patients with ICD-9 diagnosis codes representing any other chronic liver diseases and/or alcohol abuse/dependence were excluded. Using the Compensation and Pension Records Interchange (CAPRI), each corresponding electronic medical record was reviewed for the actual presence of NAFLD (determined by radiologic imaging revealing hepatic steatosis), other chronic liver diseases, and a history of alcohol abuse/dependence.
Results: Of the 17,731 veterans who had an ICD-9 discharge or encounter diagnosis of NAFLD and without ICD-9 codes representing other chronic liver diseases or alcohol abuse/dependence, medical records of a randomly chosen 5.0% (n=888) were reviewed. On chart review, 757 (85.2%) of these veterans had computed tomography or ultrasound imaging describing hepatic steatosis and no history of other chronic liver diseases or alcohol abuse/dependence. 43 (5.7%) veterans with NAFLD and no history of other chronic liver diseases or alcohol abuse/dependence used medications known to cause hepatic steatosis. In contrast, 131 (14.8%) veterans without ICD-9 codes representing other chronic liver diseases or alcohol abuse/dependence had a history documented in the medical record of other etiologies of chronic liver disease; including autoimmune hepatitis (n=1, 0.001%), hepatitis B viral infection (n=1, 0.001%), hepatitis C viral infection (n=1, 0.001%), and alcohol abuse/dependence (n=128, 14.4%).
Conclusion: Use of ICD-9 diagnosis codes to determine the presence of NAFLD and the absence of other etiologies of chronic liver diseases in VA administrative data is predictive of the presence of NAFLD without other chronic liver diseases as documented in medical records. A history of alcohol abuse/dependence is the most common alternate etiology of chronic liver disease not captured by VA administrative data. ICD-9 diagnosis codes can be reliably used for clinical outcomes research in NAFLD.
Back to 2014 Annual Meeting Posters
|
